فهرست نشریات علمی دسترسی آزاد ایران

This research introduces Serpentron, a novel snake-like robot with three joints and four links, designed for superior maneuverability in challenging environments. By extending the joint range to 180 degrees and incorporating virtual base rotation, Serpentron achieves a workspace diameter of 0.9 meters, a 50% improvement over traditional designs. A Neural Network-based Proportional-Derivative (NN-PD) controller is developed, dynamically tuning gains to track serpentine trajectories with joint angle errors below 0.02 radians and torques under 5 Nm, even under complex disturbances including Gaussian noise, velocity-dependent friction, and obstacle interactions. Simulations across single-direction, multiple-direction, and movable-direction scenarios demonstrate Serpentron’s adaptability, from planar navigation to dynamic base motion. The NN-PD controller reduces error variance by 50% compared to fixed-gain methods, leveraging a multilayer perceptron for real-time gain tuning. Torque analysis confirms the controller’s robustness against environmental uncertainties, ensuring stable performance within motor limits. Serpentron’s hyper-redundant kinematics and base mobility enable applications in confined-space exploration and volumetric inspection. Simulation results validate the model and control strategy, laying a scalable foundation for a cost-effective prototype using NVIDIA Jetson Nano and Dynamixel XL430 motors, with lightweight aluminum and ABS materials, to bridge the gap between simulation and real-world deployment in unstructured settings.